Now subclassPaths rule depends on a declaration of classifier,
which allows to track justifications dependency and invalidate
rule on incremental typechecking.
Similarly, rule for declaration now depends on all super classifiers.
Relates to machinery introduced in MPSCR-47.
Previously parent match was maintained in Controller on Java call stack.
Now this stack of parent matches is maintained manually in MatchJournal.
So MatchJournal gains new invariant: parentChunk()
returns correct parent after each logMatch() call.
Several tests fail, also need to maintain this
invariant after logActivation() calls.
Modified TYPECHECK query to include "checkingScope" query parameter:
it is a root node for which program is generated (for local typechecking it's local node).
Rules additionally match on a constraint `checkDecls(node<>)` bearing this node,
so while rules will be generated for all scopes, only those in applicable scopes will run.
Also split rule for EnumConstantDeclaration on declration checking and just typing rule.
Now primitive types in conversion relation are handled before processing of free inference vars.
So primitive types first get boxed and only then these terms can be unified with inference vars.
Fixes typechecking tests.
Type var bounds are produced together with genDecls at program start.
Now bound of type var is captured only when needed --
inside convertsTo check, leaving typeVarBound intact.
Since it's no longer possible to have cyclic terms,
the term for type variable type has to be refactored:
the bound information is attached via typeVarBound constraint
on the type.
Replace node-ptr<> type with node<> where applicable.
Use node<> type is ok in constraint arguments.
Replace instances of LateExpressionItem $() with ExpressionItem.
SNode is supported now directly as an argument to constraint/predicate.
Pointer operation is no longer necessary, as is node-ptr<> type for logical vars.
TargetSpec is to be reduced to a reference to logical variable,
simplifying the feedback predicate.
When a node can be specified directly, use target{} block.
To avoid confusion, a rule defines "target ref", which is supposed
to indicate the source code location for possible errors.
The origin node is to be referred as "origin" as before.
Temporarily rename OriginSpec's alias to "target" to reflect change
in semantics.
Temprorarily use classifier type String for internal string type.
Restore eq predicate in converts_ground_id body.
Remove origin{} block from var decl, doubtful idea.
Introduce SessionData, allow to keep arbitrary data with the session.
Extend RulesListTable with support for invalidation based on rule unique tags.
Introduce UpdatesRecorder to keep track of invalidated nodes.
Extend ApplyTemplatesStep with invalidate logic for typechecking.
Simplify design, drop unnecessary fields, minor refactorings.
Introduced a variation of conversion relation with a bit stricter boxing, needed for eq/neq.
e.g. suppose { (Number n, float f) => n == f } here f shouldn't be boxed.
Everywhere where we don't care whether it's <String> or <string> it will be handled just as String.
Where we need to discern between the cases (as in Recover), 'stringType()' dataform can be matched first.
Remove rules which handle capture+inference var case separately: it's already handled.
Remove rule with error on bound which shouldn't be triggered with new hasBound semantics.
Add symmetric rule for case value+inference var: types can be inferred nevermind relation direction.
Add rules for free vars without bounds (these ain't inference vars). Need in such vars arise when
e.g. params in classifierType terms must be somehow accessed. It's used in MethodCall rules.
During investigation a different issue has been found.
Currently, e.g. when checking args to a method call, if there's an error then inference variable remains free.
It allows to continue typechecking without spurious additional errors triggered by former ones,
but with this approach all relations must gracefully handle free vars, and it's more difficult to avoid "no type".
There's an alternative to always assign to inference var a type at hand, even it doesn't pass checks
-- but then there're more unneccessary errors. Both approaches seem valid, no action is taken now.
Enabling it with working reactivation causes spurious errors on method calls
because now it's impossible to specify feedback target for such errors fired
from reactivated rules.
The inferred type can be either capture or type variable.
This logic is covered by a new rule 'inferFromCapture'.
For examples see 'Bounds' test sample.
This inference is used in call arguments type inference.
It happens on cases when bounds arise from several places:
(a) bounds for inference variables from type vars declared in method signature (hasBound constraint)
(b) bounds of type vars used as arguments to call (typeVariableType term wiht bounds inside)
(c) bounds from captures on type variables from case (b), which come in the form of 'captureOf' term
Errors are now reported on ClassifierType instead of type argument, because some checks must be done
in captures and there's no proper way to specify origin of arg here.
Grigorii Kirgizov